setwd("C:\\Users\\aniverb\\Documents\\Grad_School\\JHU\\700")
tv=read.table("trades_pivot_2016-11-18.txt", sep="\t", header=T)
pv=read.table("trades_pivot_vol_2016-11-18.txt", sep="\t", header = T)
trade_data=read.table("trades_count_regression_2016-11-18.txt", sep="\t", header=T)
tv_pv_comb=cbind(tv[,1:16], pv[,4:16], tv[17], pv[17])
tv_pv_comb["Day"]=trade_data["Day"]
head(tv_pv_comb)
## ProductName Maturity Date X6 X7 X8 X9 X10 X11
## 1 6E 2016-06-13 2016-05-02 4743 4371 14131 19188 26520 10183
## 2 6E 2016-06-13 2016-05-03 7761 9281 23010 18470 27603 18569
## 3 6E 2016-06-13 2016-05-04 4065 10092 31111 16142 41140 12268
## 4 6E 2016-06-13 2016-05-05 11257 9318 17680 22189 30305 12814
## 5 6E 2016-06-13 2016-05-06 3582 4173 58424 29355 16648 7940
## 6 6E 2016-06-13 2016-05-09 4806 9660 12547 9459 14289 6987
## X12 X13 X14 X15 X16 X17 X18 X6.1 X7.1 X8.1 X9.1 X10.1
## 1 9504 5228 6112 2577 2616 0 0 4.195 2.388 5.414 15.202 10.399
## 2 8871 10705 7300 3884 2149 34 9 2.845 4.866 5.204 6.455 12.777
## 3 4785 4985 6424 2846 1784 15 0 2.709 5.932 11.152 7.811 9.139
## 4 7126 6431 4706 3802 1169 1 0 3.966 3.955 4.532 8.688 7.442
## 5 5493 6323 10189 4226 1304 1 0 3.432 1.998 19.902 8.070 8.105
## 6 0 0 0 0 0 16 0 2.251 4.298 6.758 4.963 7.391
## X11.1 X12.1 X13.1 X14.1 X15.1 X16.1 X17.1 X18.1 DayTradeTotal
## 1 2.440 6.733 3.743 4.343 1.836 5.190 NA NA 105173
## 2 3.812 6.492 2.912 6.776 1.797 3.144 0.866 0.837 137646
## 3 4.576 2.625 3.157 3.281 2.045 2.589 0.354 NA 135657
## 4 3.817 4.305 3.403 2.657 1.482 1.047 NA NA 126798
## 5 2.980 2.001 5.503 4.956 1.861 1.606 NA NA 147658
## 6 3.536 NA NA NA NA NA 0.635 NA 57764
## DayVolatility Day
## 1 16.844 1
## 2 23.703 2
## 3 12.542 3
## 4 14.243 4
## 5 17.925 5
## 6 8.338 6
pvals=c()
prod=c()
products=as.vector(unique(tv_pv_comb$ProductName))
significant=list(c("ZQ", "2016-08-31"), c("YM", "2017-03-17"), c("HO", "2016-11-30"), c("ZF", "2016-12-30"), c("RB", "2016-07-29"))
dfInt=data.frame(Products=c("HO", "6E", "CL", "ES", "KE", "ZL", "CL", "GC", "RB", "GE", "GE", "GE", "UB", "ZB", "ZM", "ZQ"), Maturities=c("2016-09-30","2017-06-19", "2016-07-20", "2016-09-16", "2016-12-14","2016-12-14", "2017-01-20", "2016-10-27", "2016-07-29", "2016-09-19", "2016-07-18", "2016-11-14", "2016-09-21", "2016-09-21", "2016-10-14", "2016-07-29"))
inDF=function(x, product, maturity){
if ((x[1]==product) & (x[2]==maturity)){
return(TRUE)
}else{
return(FALSE)
}
}
for (product in products){ #in results, put 3 plots on 1 line
tv_pv_combS=subset(tv_pv_comb, ProductName==product)
maturities=as.vector(unique(tv_pv_combS$Maturity))
for (maturity in maturities){
tv_pv_combM=subset(tv_pv_combS, Maturity==maturity)
if (nrow(tv_pv_combM)>=30){
cors=apply(tv_pv_combM, 1, function(x){cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use="na.or.complete")})
dfCors=data.frame(tv_pv_combM[c("Date", "Day", "DayTradeTotal")], cors=as.vector(cors))
tradeQuant=quantile(dfCors$DayTradeTotal, c(.6, .9))
tv_pv_combMQ=dfCors[(dfCors$DayTradeTotal>tradeQuant[1]) & (dfCors$DayTradeTotal<tradeQuant[2]),] #selecting high volume trades
mod=lm(tv_pv_combMQ$cors~tv_pv_combMQ$Day)
cor_hat=predict(mod, tv_pv_combMQ["Day"])
tv_pv_combMQ["cor_hat"]=cor_hat
if ((product==significant[[1]][1]) & (maturity==significant[[1]][2]) | (product==significant[[2]][1]) & (maturity==significant[[2]][2]) | (product==significant[[3]][1]) & (maturity==significant[[3]][2])| (product==significant[[4]][1]) & (maturity==significant[[4]][2])| (product==significant[[5]][1]) & (maturity==significant[[5]][2])){
png(paste("Coeff_fit", product, maturity, ".png", sep=""), width = 6, height = 4, units = 'in', res = 400)
plot(tv_pv_combMQ$Day, tv_pv_combMQ$cors, ylim=c(-1,1), xlab="Day", ylab="correlation coefficient", main=paste(product, "Maturity:", maturity), xaxt="n")
lines(tv_pv_combMQ$Day, tv_pv_combMQ$cor_hat, col="red")
at=seq(min(tv_pv_combMQ$Day), max(tv_pv_combMQ$Day), 1)
labs=rep(NA, length(at))
labs[which(at%in%tv_pv_combMQ$Day)]=as.vector(tv_pv_combMQ$Date)
axis(1, at=at, labels=labs, tick=F)
dev.off()
}
plot(tv_pv_combMQ$Day, tv_pv_combMQ$cors, ylim=c(-1,1), xlab="Day", ylab="correlation coefficient", main=paste(product, "Maturity:", maturity), xaxt="n")
lines(tv_pv_combMQ$Day, tv_pv_combMQ$cor_hat, col="red")
at=seq(min(tv_pv_combMQ$Day), max(tv_pv_combMQ$Day), 1)
labs=rep(NA, length(at))
labs[which(at%in%tv_pv_combMQ$Day)]=as.vector(tv_pv_combMQ$Date)
axis(1, at=at, labels=labs, tick=F)
stat=summary(mod)
stat=summary(mod)
cof=stat$coefficients
pval=cof[,4][2]
pvals=append(pvals, pval, after=length(pvals))
prod=append(prod, paste(product, maturity), after=length(prod))
trades=as.vector(as.matrix(t(tv_pv_combM[,4:16])))
priceVol=as.vector(as.matrix(t(tv_pv_combM[,17:29])))
df=data.frame(trades, priceVol)[order(trades),]
df=na.omit(df)
n=nrow(df)
dfTrim=df[ceiling(.05*n):round(.95*n),]
dfTrimO=dfTrim[order(dfTrim$priceVol),]
n2=nrow(dfTrimO)
dfTrimP=dfTrimO[ceiling(.05*n2):round(.95*n2),]
plot(dfTrimP, xlab="trade volume hourly", ylab="price volatility hourly", main=paste(product, "Maturity:", maturity), pch=20)
daily=tv_pv_combM[,30:31][order(tv_pv_combM$DayTradeTotal),]
daily=na.omit(daily)
nD=nrow(daily)
dailyTrim=daily[ceiling(.05*nD):round(.95*nD),]
dailyTrimO=dailyTrim[order(dailyTrim$DayVolatility),]
nD2=nrow(dailyTrimO)
dailyTrimP=dailyTrimO[ceiling(.05*nD2):round(.95*nD2),]
plot(dailyTrimP, xlab="trade volume daily", ylab="price volatility daily", main=paste(product, "Maturity:", maturity), pch=20)
t=apply(dfInt, 1, inDF, product, maturity)
if (TRUE %in% t){
png(paste(product, maturity, "hourly.png", sep=""), width = 6, height = 4, units = 'in', res = 300)
plot(dfTrimP, xlab="trade volume hourly", ylab="price volatility hourly", main=paste(product, "Maturity:", maturity), col="pink3", pch=20)
dev.off()
png(paste(product, maturity, "daily.png", sep=""), width = 6, height = 4, units = 'in', res = 300)
plot(dailyTrimP, xlab="trade volume daily", ylab="price volatility daily", main=paste(product, "Maturity:", maturity), col="pink3", pch=20)
dev.off()
}
}
}
}
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
## Warning in cor(as.numeric(x[4:16]), as.numeric(x[17:29]), use =
## "na.or.complete"): the standard deviation is zero
pvalsO=order(pvals)
corPvals=data.frame(prod, pvals)[pvalsO,]
nPvals=nrow(corPvals)
nPvals
## [1] 126
corPvals["id"]=1:nPvals #B-H Procedure
corPvals["q"]=nPvals*corPvals$pvals/corPvals$id
f=c()
for(i in 1:nPvals){f=append(f, min(corPvals$q[i:nPvals]), length(f))}
corPvals["f"]=f
corSig=corPvals[which(corPvals$pvals<.05),]
nSig=nrow(corSig)
nSig
## [1] 16
corSig #5 rejected at FDR=.20
## prod pvals id q f
## 111 ZQ 2016-08-31 0.001326200 1 0.1671012 0.1181514
## 78 YM 2017-03-17 0.002226816 2 0.1402894 0.1181514
## 36 HO 2016-11-30 0.002813129 3 0.1181514 0.1181514
## 90 ZF 2016-12-30 0.004494289 4 0.1415701 0.1415701
## 59 RB 2016-07-29 0.006932494 5 0.1746988 0.1746988
## 43 KE 2017-05-12 0.014340904 6 0.3011590 0.3011590
## 6 CL 2016-07-20 0.018004212 7 0.3240758 0.3240758
## 2 6E 2016-12-19 0.021064406 8 0.3317644 0.3302329
## 103 ZM 2016-12-14 0.023588068 9 0.3302329 0.3302329
## 32 HO 2016-07-29 0.027519890 10 0.3467506 0.3467506
## 24 GC 2017-04-26 0.032399027 11 0.3711161 0.3546173
## 29 GE 2016-12-19 0.034721438 12 0.3645751 0.3546173
## 53 NG 2017-04-26 0.039869811 13 0.3864305 0.3546173
## 3 6E 2017-03-13 0.040719034 14 0.3664713 0.3546173
## 115 ZQ 2016-12-30 0.042216346 15 0.3546173 0.3546173
## 55 NQ 2016-09-16 0.047810841 16 0.3765104 0.3765104
#trimming seems to not have worked for ZQ and RB since they are close to maturity
set.seed(3-3-17)
u=sort(runif(length(pvals)))
#png("unifQQ.png")
plot(u, corPvals$pvals, ylab="p-values", xlab="uniform", main="Q-Q Plot")
abline(0,1, col="red")
#dev.off()